Not so fast Obama supporters. The race is not over in New Mexico.

True, if you’re reasonably informed, you probably wouldn’t bet on any Republican being elected to office this year—GOP candidates can thank George W. Bush and the economy for that—but when it comes to presidential races in New Mexico, we are a capricious state.

Back in 2004, conventional wisdom held that winning in the urban areas of New Mexico meant winning the state. The presidential candidate who carries Albuquerque, Santa Fe and Las Cruces wins. The Bush campaign, however, redefined this winning strategy by campaigning in rural areas where no president had campaigned before, and in the end, the enormous turnout and support he got in Lea, Curry and San Juan counties gave him a less than 1 percent margin of victory. In short, Farmington did it for Bush in 2004.

And it could happen again. One last-minute trip to San Juan County by Sarah Palin could have an enormous effect in tightening the race. Overall, it doesn’t seem like the McCain-Palin ticket campaigned in the rural areas as much as Bush-Cheney did four years ago, but it is still possible that turnout in these areas will be huge for the GOP. More to the point, if McCain’s going to keep New Mexico, it needs to be.

Our cities are embracing Obama for certain. If the early and absentee voting trends continue through Election Day, this year’s Democratic presidential candidate could earn much better numbers than John Kerry did four years ago. For example, Kerry won Bernalillo County by approximately 10,000, but lost the state by about 6,000 votes. (I’m going by memory because the Secretary of State’s Web site does not have this information posted. The homepage says past election results are being updated “to fit the new look and feel” of the Web site. Bad timing, Mary!)

The good news for Obama, considering the groundswell of support evidenced at the University of New Mexico rally last weekend, is that it seems possible he could win Bernalillo by 20,000 votes. If that happens, the rural vote might not matter as much.

Also, despite the infinite wisdom of Fernando C de Baca, the real question is, will Hispanics vote as a bloc this year, as in will 65 percent or more of them support the Democrat? Bush received 39 percent of the Hispanic vote in 2004. If McCain can split the Hispanic vote along these same lines, he may still be in the game.

News is circulating that the pro-life political wing of the Catholic Church plans to leaflet cars in the state on Sunday with the same “vote GOP” emotional appeal that had a positive impact on Bush’s numbers when the same strategy was employed two days before the election in 2004.

Lastly, while Democrats proportionately outnumbered Republicans in early and absentee voting by a 2-1 margin as of yesterday, this does not guarantee the trend will continue on Election Day.

And while you consider this, pay careful attention to the poll in Sunday’s Albuquerque Journal. Local pollster Brian Sanderoff knows how to poll our state better than anyone, and his last poll of the season is always the one to watch. In understanding how to gauge the veracity of polls you see online, keep in mind that the larger the sample, the more reliable the poll and the smaller the margin of error. Sanderoff’s final poll will sample more than 1,000 people, which is larger than any of the samples used in other polls posted on the Internet. Sanderoff’s staff conducts person-to-person interview that are more reliable than the automated phone polling data folks like www.realclearpolitics.com depend on.

This final poll might not be spot-on—polling is not an exact science—but it should hint to whether we are in for a late night.